Background. Prognostication after cardiac arrest in children is challenging due to a lack of validated methods to evaluate direct brain injury. The objective of this multicenter study was to analyze biomarker accuracy to prognosticate outcome 1 year post-arrest.
Methods. Fourteen U.S. centers enrolled 164 children ages 48 h - 17 years with pre-arrest Pediatric Cerebral Performance Category score of 1-3 who were admitted to an intensive care unit after cardiac arrest. Glial fibrillary acidic protein (GFAP), ubiquitin carboxyl-terminal esterase- L1 (UCH-L1), neurofilament light (NfL), and Tau concentrations were measured in blood samples from post-arrest days 1-3 using Quanterix Simoa 4-Plex assay, Billerica, MA. Unfavorable outcome was death or survival with Vineland Adaptive Behavioral Scale-Third Edition score < 70 at 1 year. We analyzed area under receiver operator curve (AUROC) and performed multivariate logistic regressions to determine the association of each biomarker with outcome on days 1-3.
Results. Fifty of 120 children with primary outcomes available had an unfavorable outcome, including 43 deaths. Compared to those with favorable outcomes, more children with unfavorable outcome had out-of-hospital (36% vs. 70%) and unwitnessed (7% vs. 46%) events, p<0.05. For days 1-3, concentrations of all four measured biomarkers were increased in children with an unfavorable vs. favorable outcome, p<0.05. On post-arrest day 1, NfL demonstrated the best outcome classification (AUROC 0.731 [95% confidence interval 0.642, 0.820]) while UCH-L1 performed best on days 2 (0.860 [0.785, 0.935]) and 3 (0.837 [0.747, 0.926]). After covariate adjustment, NfL concentrations on day 1 (odds ratio 5.9 [95% confidence interval 1.8, 19.2], day 2 (11.9 [3.8, 36.9]), and day 3 (10.2 [3.1, 33.3]), UCH-L1 on day 2 (11.3 [3.0, 42.4]) and day 3 (7.6 [2.1, 27.1]), GFAP on day 2 (2.3 [1.2, 4.5]) and day 3 (2.2 [1.2, 4.0]), and tau on day 1 (2.4 [1.1, 5.3]), day 2 (2.3 [1.3, 4.0]), and day 3 (2.0 [1.2, 3.6]) were associated with unfavorable outcome, p<0.05.
Conclusions. Blood-based brain injury biomarkers accurately prognosticated death or unfavorable adaptive behavior composite outcome at 1 year after pediatric cardiac arrest. Accuracy of biomarkers to predict neurodevelopmental outcomes beyond 1 year should be evaluated.